This skill serves as a comprehensive framework for developing agentic AI systems, focusing on reliability, observability, and graceful failure. It provides standardized implementations for ReAct loops, plan-and-execute architectures, secure tool integration, and sophisticated memory systems including short-term, long-term, and working memory. By applying these patterns, developers can transition from simple LLM calls to robust, autonomous workflows capable of handling complex multi-step tasks while maintaining safety guardrails and multi-agent coordination.
Key Features
01ReAct (Reasoning + Acting) loop implementation patterns
02Plan-and-Execute architectures for complex multi-step tasks
03Hierarchical multi-agent orchestration patterns
04Secure tool definition and sandboxed execution frameworks
05Multi-layered memory systems (Short-term, Vector-based, and Working)
061 GitHub stars
